Performance Evaluation for AVS2 Scene Video Coding Techniques

摘要

Different from conventional video coding techniques, the emerging AVS2 standard (the second generation of Audio Video coding Standard, AVS2 for short) introduces some new coding techniques mainly focuses on video coding on scene videos. Scene videos are videos with limited scenes, such as surveillance videos, conference videos, etc. For scene videos, due to the limited scenes, great redundancy especially in background region is retained in the video pictures. AVS2 scene video coding techniques aim at significantly reducing the redundancy so as to achieve higher compression performance. This paper introduces the main coding techniques in AVS2 and demonstrates the coding performance. Experimental results show that the coding efficiency has almost doubled over the main group on scene videos which brings new opportunities to the related industries on specific applications.